Silex launches EP-200N SoM for secure 2-TOPS edge video processing

The shift toward local AI inference requires hardware that can process complex video pipelines without the latency or cost of cloud computing. By providing a production-ready module with integrated security and safety cores, Silex addresses the fragmented hardware landscape for edge-based video analytics. This enables streaming and surveillance applications to handle 4K30 feeds or multi-camera sensor fusion closer to the source, reducing backhaul bandwidth requirements. As the market moves toward hybrid architectures, watch the adoption rate of these modules in specialized verticals like medical remote monitoring and industrial robotics where deterministic performance is non-negotiable.

The launch of the EP-200N coincides with the broader commercial availability of NXP’s i.MX 95 series, which NXP targeted for broad silicon availability in May 2026. The processor represents the highest-performance tier in NXP's i.MX line, featuring the eIQ Neutron Neural Processing Unit (NPU). According to NXP reports from April 2026, the Neutron NPU offers up to 4x the machine learning performance and 2x the data throughput of previous generations like the i.MX 8M Plus, while improving overall power efficiency by roughly 15%.

Silex is competing in an increasingly crowded System-on-Module (SoM) ecosystem focused on high-performance edge AI. In late 2025 and 2026, other hardware providers also began shipping competing i.MX95 solutions. For instance, Ezurio (formerly Laird Connectivity) announced its Nitrogen95 SoM in December 2025, which utilizes the SMARC 2.2 form factor, while Advantech introduced its AOM-5521 and AOM-2521 modules in August 2025. These hardware launches reflect a wider industry pivot toward localized video processing, as IDC projected in May 2025 that global edge computing spending would reach $261 billion by the end of the year.

Technological partnerships are also deepening to simplify the deployment of these AI-heavy chips. In April 2026, Silex announced a collaboration with Edge Impulse to integrate MLOps tools directly onto their modules, specifically for vision-guided robotics and smart manufacturing. This software-hardware synergy is critical as video processing platforms are expected to reach a market value of $38.85 billion by 2035, per Roots Analysis (May 2025), driven by the demand for real-time analytics and energy-efficient encoding at the device level.